检索结果-内蒙古大学图书馆

Cross-Domain Motion Transfer via Safety-Aware Shared Latent Space Modeling

IEEE robotics AND automation LETTERS 2020年第2期5卷 2634-2641页

作者： Choi, Sungjoon Kim, Joohyung Disney Res Glendale CA 91201 USA Univ Illinois Champaign IL 61820 USA

This letter presents a data-driven motion retargeting method with safety considerations. In particular, we focus on handling self-collisions while transferring poses between different domains. To this end, we first propose leveraged Wasserstein auto-encoders (LWAE) which leverage both positive and negative data where negative data consist of self-collided poses. Then, we extend this idea to multiple domains to have a shared latent space to perform motion retargeting. We also present an effective self-collision handling method based on solving inverse kinematics with augmented targets that is used to collect collision-free poses. The proposed method is extensively evaluated in a diverse set of motions from human subjects and an animation character where we show that incorporating negative data dramatically reduces self-collisions while preserving the quality of the original motion.

关键词： deep learning in robotics and automation motion and path planning collision avoidance

来源：评论

学校读者我要写书评

暂无评论

Addressing the Sim2Real Gap in Robotic 3-D Object Classification

引用

IEEE robotics AND automation LETTERS 2020年第2期5卷 407-413页

作者： Weibel, Jean-Baptiste Patten, Timothy Vincze, Markus TU Wien Automat & Control Inst Vis Robot Lab A-1040 Vienna Austria

Object classification with 3D data is an essential component of any scene understanding method. It has gained significant interest in a variety of communities, most notably in robotics and computer graphics. While the advent of deep learning has progressed the field of 3D object classification, most work using this data type are solely evaluated on CAD model datasets. Consequently, current work does not address the discrepancies existing between real and artificial data. In this work, we examine this gap in an indoor service robotic context by specifically addressing the problem of classification when transferring from artificial CAD models to real reconstructed objects. This is performed by training on ModelNet (CAD models) and evaluating on ScanNet (objects extracted from reconstructed rooms). We show that standard methods do not perform well in this task. We thus introduce a method that carefully samples object parts that are reproducible under various transformations and hence robust. Using graph convolution to classify the composed graph of parts, our method improves upon the baseline. Code is publicly available at https://***/***/cad2real_object_clf.

关键词： Object detection segmentation and categorization deep learning in robotics and automation RGB-D perception

来源：评论

学校读者我要写书评

暂无评论

Safe and Fast Tracking on a Robot Manipulator: Robust MPC and Neural Network Control

引用

IEEE robotics AND automation LETTERS 2020年第2期5卷 3050-3057页

作者： Nubert, Julian Koehler, Johannes Berenz, Vincent Allgoewer, Frank Trimpe, Sebastian Max Planck Inst Intelligent Syst Intelligent Control Syst Grp D-70569 Stuttgart Germany Swiss Fed Inst Technol CH-8092 Zurich Switzerland Univ Stuttgart Inst Syst Theory & Automat Control D-70550 Stuttgart Germany Max Planck Inst Intelligent Syst Autonomous Mot Dept D-72076 Tubingen Germany

Fast feedback control and safety guarantees are essential in modern robotics. We present an approach that achieves both by combining novel robust model predictive control (MPC) with function approximation via (deep) neural networks (NNs). The result is a new approach for complex tasks with nonlinear, uncertain, and constrained dynamics as are common in robotics. Specifically, we leverage recent results in MPC research to propose a new robust setpoint tracking MPC algorithm, which achieves reliable and safe tracking of a dynamic setpoint while guaranteeing stability and constraint satisfaction. The presented robust MPC scheme constitutes a one-layer approach that unifies the often separated planning and control layers, by directly computing the control command based on a reference and possibly obstacle positions. As a separate contribution, we show how the computation time of the MPC can be drastically reduced by approximating the MPC law with a NN controller. The NN is trained and validated from offline samples of the MPC, yielding statistical guarantees, and used in lieu thereof at run time. Our experiments on a state-of-the-art robot manipulator are the first to show that both the proposed robust and approximate MPC schemes scale to real-world robotic systems.

关键词： Robustness Electron tubes Safety Manipulators Task analysis Uncertainty deep learning in robotics and automation motion control optimization and optimal control redundant robots robust adaptive control of robotic systems

来源：评论

学校读者我要写书评

暂无评论

iART: learning From Demonstration for Assisted Robotic Therapy Using LSTM

引用

IEEE robotics AND automation LETTERS 2020年第2期5卷 477-484页

作者： Pareek, Shrey Kesavadas, Thenkurussi Univ Illinois Dept Ind & Enterprise Syst Engn Champaign IL 61820 USA

In this letter, we present an intelligent Assistant for Robotic Therapy (iART), that provides robotic assistance during 3D trajectory tracking tasks. We propose a novel LSTM-based robot learning from demonstration (LfD) paradigm to mimic a therapist's assistance behavior. iART presents a trajectory agnostic LfD routine that can generalize learned behavior from a single trajectory to any 3D shape. Once the therapist's behavior has been learned, iART enables the patient to modify this behavior as per their preference. The system requires only a single demonstration of 2 minutes and exhibits a mean accuracy of 91.41% in predicting, and hence mimicking a therapist's assistance behavior. The system delivers stable assistance in realtime and successfully reproduces different types of assistance behaviors.

关键词： Rehabilitation robotics deep learning in robotics and automation learning from demonstration AI-based methods

来源：评论

学校读者我要写书评

暂无评论

OriNet: Robust 3-D Orientation Estimation With a Single Particular IMU

引用

IEEE robotics AND automation LETTERS 2020年第2期5卷 399-406页

作者： Esfahani, Mahdi Abolfazli Wang, Han Wu, Keyu Yuan, Shenghai Nanyang Technol Univ Sch Elect & Elect Engn Singapore 639798 Singapore

Estimating the robot's heading is a crucial requirement in odometry systems which are attempting to estimate the movement trajectory of a robot. Small errors in the orientation estimation result in a significant difference between the estimated and real trajectory, and failure of the odometry system. The odometry problem becomes much more complicated for micro flying robots since they cannot carry massive sensors. In this manner, they should benefit from the small size and low-cost sensors, such as IMU, to solve the odometry problem, and industries always look for such solutions. However, IMU suffers from bias and measurement noise, which makes the problem of position and orientation estimation challenging to be solved by a single IMU. While there are numerous studies on the fusion of IMU with other sensors, this study illustrates the power of the first deep learning framework for estimating the full 3D orientation of the flying robots (as yaw, pitch, and roll in quaternion coordinates) accurately with the presence of a single IMU. A particular IMU should be utilized during the training and testing of the proposed system. Besides, a method based on the Genetic Algorithm is introduced to measure the IMU bias in each execution. The results show that the proposed method improved the flying robots' ability to estimate their orientation displacement by approximately 80% with the presence of a single particular IMU. The proposed approach also outperforms existing solutions that utilize a monocular camera and IMU simultaneously by approximately 30%.

关键词： Localization SLAM deep learning in robotics and automation autonomous vehicle navigation

来源：评论

学校读者我要写书评

暂无评论

A General Framework to Increase Safety of learning Algorithms for Dynamical Systems Based on Region of Attraction Estimation

引用

IEEE TRANSACTIONS ON robotics 2020年第5期36卷 1472-1490页

作者： Zhou, Zhehua Oguz, Ozgur S. Leibold, Marion Buss, Martin Tech Univ Munich Chair Automat Control Engn D-80290 Munich Germany

Although the state-of-the-art learning approaches exhibit impressive results for dynamical systems, only a few applications on real physical systems have been presented. One major impediment is that the intermediate policy during the training procedure may result in behaviors that are not only harmful to the system itself but also to the environment. In essence, imposing safety guarantees for learning algorithms is vital for autonomous systems acting in the real world. In this article, we propose a computationally effective and general safe learning framework, specifically for complex dynamical systems. With a proper definition of the safe region, a supervisory control strategy, which switches the actions applied on the system between the learning-based controller and a predefined corrective controller, is given. A simplified system facilitates the estimation of the safe region for the high-dimensional dynamical system. During the learning phase, the belief of the safe region is updated with the actual execution results of the corrective controller, which in turn enables the learning-based controller to have more freedom in choosing its actions. Two examples are given to demonstrate the performance of the proposed framework, one simple inverted pendulum to illustrate the online adaptation method, and one quadcopter control task to show the overall performance.

关键词： Safety Dynamical systems Heuristic algorithms Reinforcement learning Control systems Task analysis Estimation deep learning in robotics and automation learning and adaptive systems robot safety safe reinforcement learning

来源：评论

学校读者我要写书评

暂无评论

CNN Based Road User Detection Using the 3D Radar Cube

引用

IEEE robotics AND automation LETTERS 2020年第2期5卷 1263-1270页

作者： Palffy, Andras Dong, Jiaao Kooij, Julian F. P. Gavrila, Dariu M. Delft Univ Technol Intelligent Vehicles Grp NL-2628 Delft Netherlands

This letter presents a novel radar based, single-frame, multi-class detection method for moving road users (pedestrian, cyclist, car), which utilizes low-level radar cube data. The method provides class information both on the radar target- and object-level. Radar targets are classified individually after extending the target features with a cropped block of the 3D radar cube around their positions, thereby capturing the motion of moving parts in the local velocity distribution. A Convolutional Neural Network (CNN) is proposed for this classification step. Afterwards, object proposals are generated with a clustering step, which not only considers the radar targets;positions and velocities, but their calculated class scores as well. In experiments on a real-life dataset we demonstrate that our method outperforms the state-of-the-art methods both target- and object-wise by reaching an average of 0.70 (baseline: 0.68) target-wise and 0.56 (baseline: 0.48) object-wise F1 score. Furthermore, we examine the importance of the used features in an ablation study.

关键词： Object detection segmentation and categorization sensor fusion deep learning in robotics and automation

来源：评论

学校读者我要写书评

暂无评论

Self-Supervised Linear Motion Deblurring

引用

IEEE robotics AND automation LETTERS 2020年第2期5卷 2475-2482页

作者： Liu, Peidong Janai, Joel Pollefeys, Marc Sattler, Torsten Geiger, Andreas Swiss Fed Inst Technol Comp Vis & Geometry Grp Dept Comp Sci CH-8092 Zurich Switzerland Max Planck Inst Intelligent Syst Autonomous Vis Grp D-41296 Tubingen Germany Microsoft Mixed Real & Artificial Intelligence La CH-8001 Zurich Switzerland Chalmers Univ Technol Comp Vis & Med Image Anal Grp S-41296 Gothenburg Sweden

Motion blurry images challenge many computer vision algorithms, e.g., feature detection, motion estimation, or object recognition. deep convolutional neural networks are state-of-the-art for image deblurring. However, obtaining training data with corresponding sharp and blurry image pairs can be difficult. In this letter, we present a differentiable reblur model for self-supervised motion deblurring, which enables the network to learn from real-world blurry image sequences without relying on sharp images for supervision. Our key insight is that motion cues obtained from consecutive images yield sufficient information to inform the deblurring task. We therefore formulate deblurring as an inverse rendering problem, taking into account the physical image formation process: we first predict two deblurred images from which we estimate the corresponding optical flow. Using these predictions, we re-render the blurred images and minimize the difference with respect to the original blurry inputs. We use both synthetic and real dataset for experimental evaluations. Our experiments demonstrate that self-supervised single image deblurring is really feasible and leads to visually compelling results. Both the code and datasets are available at https://***/ethliup/SelfDeblurhttps://***/ethliup/SelfDeblur.

关键词： Computer vision for automation deep learning in robotics and automation

来源：评论

学校读者我要写书评

暂无评论

RSL-Net: Localising in Satellite Images From a Radar on the Ground

引用

IEEE robotics AND automation LETTERS 2020年第2期5卷 1087-1094页

作者： Tang, Tim Yuqing De Martini, Daniele Barnes, Dan Newman, Paul Univ Oxford Oxford Robot Inst Oxford OX1 2JD England

This letter is about localising a vehicle in an overhead image using FMCW radar mounted on a ground vehicle. FMCW radar offers extraordinary promise and efficacy for vehicle localisation. It is impervious to all weather types and lighting conditions. However the complexity of the interactions between millimetre radar wave and the physical environment makes it a challenging domain. Infrastructure-free large-scale radar-based localisation is in its infancy. Typically here a map is built and suitable techniques, compatible with the nature of sensor, are brought to bear. In this work we eschew the need for a radar-based map;instead we simply use an overhead image - a resource readily available everywhere. This letter introduces a method that not only naturally deals with the complexity of the signal type but does so in the context of cross modal processing.

关键词： Autonomous vehicle navigation deep learning in robotics and automation localization range sensing

来源：评论

学校读者我要写书评

暂无评论

RLBench: The Robot learning Benchmark & learning Environment

引用

IEEE robotics AND automation LETTERS 2020年第2期5卷 3019-3026页

作者： James, Stephen Ma, Zicong Arrojo, David Rovick Davison, Andrew J. Imperial Coll London Dyson Robot Lab London SW7 2BU England Imperial Coll London UROP London SW7 2BU England

We present a challenging new benchmark and learning-environment for robot learning: RLBench. The benchmark features 100 completely unique, hand-designed tasks, ranging in difficulty from simple target reaching and door opening to longer multi-stage tasks, such as opening an oven and placing a tray in it. We provide an array of both proprioceptive observations and visual observations, which include rgb, depth, and segmentation masks from an over-the-shoulder stereo camera and an eye-in-hand monocular camera. Uniquely, each task comes with an infinite supply of demos through the use of motion planners operating on a series of waypoints given during task creation time;enabling an exciting flurry of demonstration-based learning possibilities. RLBench has been designed with scalability in mind;new tasks, along with their motion-planned demos, can be easily created and then verified by a series of tools, allowing users to submit their own tasks to the RLBench task repository. This large-scale benchmark aims to accelerate progress in a number of vision-guided manipulation research areas, including: reinforcement learning, imitation learning, multi-task learning, geometric computer vision, and in particular, few-shot learning. With the benchmark's breadth of tasks and demonstrations, we propose the first large-scale few-shot challenge in robotics. We hope that the scale and diversity of RLBench offers unparalleled research opportunities in the robot learning community and beyond. Benchmarking code and videos can be found at https://***/view/rlbench.

关键词： Task analysis Benchmark testing Cameras learning (artificial intelligence) Tools Robot vision systems learning from demonstration deep learning in robotics and automation performance evaluation and benchmarking perception for grasping and manipulation

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：